DIYHiFi.org

For the sake of audio
It is currently Mon Dec 18, 2017 8:35 pm

All times are UTC




Post new topic Reply to topic  [ 4 posts ] 
Author Message
PostPosted: Fri Aug 02, 2013 8:10 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Folks,

I recently had occasion to retrieve a "lost webpage" from the internet archive.

It had posted several files from Earl Geddes's Distortion Perception study, which gave results that are amusing.

Here is a link to an archive of all four files.

http://s000.tinyupload.com/index.php?file_id=01182935901763662869

I have anonymised them, though anyone with a modicum of google-fu can derive solution, but that would be cheating.

The file called sample_0.wav is the original (hand on my heart and all that) the others, well your best guess.

Your assignment, should you accept it, is to listen to to the three distorted samples and rank them as follows:

1) Rank by what you think is the most distorted file (subjectively)
2) Rank by which file you find most pleasing to listen to

So there is noskullduggery or such I am happy to send the solution to roiibm before anyone posts their results, to be made public after a few weeks.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Sun Aug 04, 2013 3:30 am 
Offline
Cow

Joined: Fri Jan 07, 2005 10:28 pm
Posts: 102
There's a reason why those files were removed, and why the article never passed peer review.

The reason is that when distortion is generated via a static nonlinearity in the digital domain, aliasing occurs if no oversampling is done. Earl didn't handle that situation, and has admitted as much, but I can't find the article at the pub. Keith Howard correctly pointed out the same thing in the documentation associated with his add_distortion code, used in his article, Euphonic Distortion: Naughty but Nice?.

In this case, Stereophile got it right and Geddes got it wrong. A stopped clock and all that.


Top
 Profile  
Reply with quote  
PostPosted: Mon Aug 05, 2013 6:20 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Andy,

andy_c wrote:
There's a reason why those files were removed, and why the article never passed peer review.

The reason is that when distortion is generated via a static nonlinearity in the digital domain, aliasing occurs if no oversampling is done.


First, your claim that oversampling must be used is false. The distortion must be low passed at 1/2 FS, at least in theory, but oversampling is NOT necessary.

At the recent increasing use of digital filters that release substantial aliases or even the removal of all filters at least for playback shows aliasing distortion is by far less problematic sonically than many FUD merchants have been making out.

andy_c wrote:
Earl didn't handle that situation, and has admitted as much, but I can't find the article at the pub.


I have used my best google-foo and have also come up empty. If you want to make such claims about what someone else said a reference is desired.

Secondly, even if he didn't handle this situation (and I suspect many non-linear plugins for audio workstations and DSP routines also do not do this - which is why their "tubesound" does never sound like tubes), the beauty is that while it may make his metric (which I never adopted - I still use D.E.L. Shorter's myself) invalid or not, for the topic here it matters zip.

If a given algorithm measured 0.1% THD(+N) or if it measured 9.6% THD(+N), this THD includes the aliasing and so for my little demonstration to show just how meaningless THD is, it is perfectly sufficient.

I notice that Earl has not updated his paper based on the same methodology that created these files and that it remains available on his site. So perhaps he too feels that his work drew essentially correct conclusion but is being blackballed on tecnicalities by teh usual blue meanies....

andy_c wrote:
Keith Howard correctly pointed out the same thing in the documentation associated with his add_distortion code, used in his article, Euphonic Distortion: Naughty but Nice?.

In this case, Stereophile got it right and Geddes got it wrong. A stopped clock and all that.


I am not sure what you are implying Keith Howard got right or wrong.

However, the whole subject falls under several key factors in psychoacoustics - which Keith studiously ignores in order to make his point:

1) Masking of distortion
2) Level and frequency dependency of Masking
3) The "Missing Fundamental" Effect

Now Hiraga's thesis was derived mainly empirically in a day and time when all Amplifiers had quite high levels of distortion and when those using solid state had large amounts of feedback at low frequencies but very little at high frequencies while distortion rose dramatically towards high frequencies and where re-shaping all this distortion into a "Hiraga pattern" had distinct sonic advantages.

In fact, the same still holds today, though nowadays we need much less distortion to achieve masking, for example.

In Keith Howards in my view committed a cardinal mistake in his testing by doing it without a correct level reference. As distortion perception and audibility is greatly SPL dependent - not having a calibrated SPL (the THX convention may be applied, though I personal prefer that used by German Radio & TV, though real differences are minimal) makes the results of testing THD bylistening not very useful.

From previous tests KH published I would infer he is what I call a "quiet listener", listening mostly at quite low levels, where for example the amounts of distortion he cites in his article would indeed exceed audibility thresholds.

So this would be like pairing a normal (7W @ 3% THD) 300B SE Amplifier without NFB with a BBC LS-3/5, a recipe for sonic disaster.

My personal listening preferences would go to > 98dB/1W/1m (not 2.83V) for use of a "high power" 300B Amplifier (read 16W @ 5% THD) where at the resulting 107dB peak the HD would be by far less audible.

Anyway, non of we have debated here will impact in any way on the little experiment I suggested above.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Mon Aug 05, 2013 8:20 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Gang,

Hephaestus Audio Blog:

"I should clarify that statement: Everybody outside of the audio industry realizes that this test is silly. However, in the audio industry it is considered the most somber science to conduct a double-blind (ABX) test that flips back and forth between two units under test every few seconds.

Physicist Richard Feynman had a term for this, he called it 'Cargo Cult Science'. It is the sort of science you do when you appear to go though all the right motions, but are missing the big picture. You are doing everything you are 'supposed to', but you’re not getting useful results. "

I just came across this blog, while I am not necessarily in agreement with all the conclusion drawn (and they certainly can keep their Class D Amp's), it presents a lot of good material on how we really hear and how this relates to the equipment we use to play music.

Highly recommended for anyone who reads the details and is willing to draw their own conclusions based on them, rather than skipping to the conclusions and accepting them at face value if they fit ones belief system or rejecting them if they do not...

http://hephaestusaudio.com/delphi/

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 4 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group